MANUAL VERSUS MECHANICAL ESOPHAGOGASTRIC ANASTOMOSIS AFTER RESECTION FOR CARCINOMA - A CONTROLLED TRIAL

Background. Mechanical anastomosis has been claimed to reduce the rate of leakage compared with manual anastomosis. No randomized trials hav e been performed to date to prove this specifically esophagogastric an astomosis. Methods. One hundred fifty-four patients, 139 men and 15 wo men ranging in age from 36 to 83 years (mean, 50 +/- 10 years) and und ergoing elective resection of esophageal or cardial carcinoma, were in cluded in this multiinstitutional (14 centers) randomized study compar ing the rate of anastomotic leakage after esophagogastric anastomosis performed manually or mechanically. Eligible for this study were patie nts with esophageal or cardial carcinoma located between the esophagog astric junction (included) and the upper border of the aortic arch. Th e choice between the esophagogastric junction (included) and the upper border of the aortic arch. The choice between resection with or witho ut thoracotomy was left to the discretion of the operating surgeon. Pr oximal resection of the fundus was mandatory. Intestinal tract continu ity was reestablished in a one-stage procedure by an esophagogastric a nastomosis without interposition of either the jejunum or the colon. T he site of the anastomosis could be either intrathoracic or cervical. The principal end point was anastomotic leakage as judged by (1) egree s of intestinal fluids or orally ingested methylene blue through drain s, (2) sodium diatrozate swallow prescribed either routinely for all p atients between postoperative days 3 and 8 or because of signs of leak age, or (3) reoperation or autopsy. Results. After two patients were w ithdrawn for protocol violation, 152 patients, 74 in the manual group and 78 in the mechanical group, were studied. The number of anastomoti c leakages was identical in both groups (n = 12, 16% and 15%, respecti vely). Overall 30-day mortality was 11%. Fewer deaths occurred in the manual group (7%), which had three anastomotic leakages, than in the m echanical group (15%), which had five anastomotic leakages, and fewer repeat operations were done in the manual group (n = 9) than in the me chanical group (n = 13), but both of these differences were not statis tically significant. The duration of anastomosis and of operation was similar in both groups. In the mechanical group 16 anastomoses (20%) g ave rise to technical mishaps (either in the fashioning of the purse-s tring, dilation of the esophagus, or in stapling). Among the factors r ecognized as potentially preventing leakage, only testing for airtight ness was significantly correlated with less postoperative leakage (p < 0.05). Eight postoperative strictures were recorded at 3 months in 62 (13%) patients in the manual group, whereas seven strictures occurred in 53 (13%) patients in the mechanical group. Conclusions. When mecha nical staples rather than manual sutures are chosen, the disadvantages (technical mishaps and higher costs) are not counterbalanced by a gai n of time or a decrease in the rate of severity of anastomotic leakage .