Anastomotic technique influences outcomes after partial gastrectomy for adenocarcinoma

The proper reconstructive technique after partial gastrectomy for adenocarcinoma of the stomach is often debated, but few data exist to clarify the issue. We evaluated outcomes after different anastomoses used during partial gastrectomy for gastric adenocarcinoma. We reviewed the hospital records of all 277 patients who underwent operation for gastric cancer at our institution from 1970 to 1996. Of 118 partial gastrectomies performed with curative intent 57 anastomoses were Billroth II gastrojejunostomies, 22 were Billroth I gastroduodenal reconstructions, and 39 were Roux-en-Y gastrojejunostomies. There was no difference in the incidence of early gastric emptying problems or early or late postoperative obstruction among the groups. Average hospital stay was 14 days for the Billroth I group, 15 days for those with Billroth II reconstructions, and 22 days for the Roux-en-Y cohort. Documented late gastric outlet obstruction occurred in 29 per cent of patients having Billroth I and in 33 per cent of those with Billroth II anastomoses. Antecolic anastomoses represented 30 (53 per cent) and retrocolic 27 (47 per cent) of the 57 Billroth II reconstructions performed. Late gastric outlet obstructions occurred in seven (23 per cent) patients who had antecolic reconstructions and in just one (4 per cent) with a retrocolic anastomosis (P < 0.05). Five-year cumulative survival was lower for patients having Billroth I reconstructions than for those with Billroth II (P < 0.05). Among patients with Billroth II reconstructions, 5-year cumulative survival was lower for those with antecolic reconstructions compared with those with retrocolic anastomoses (P < 0.05). Although conventional teaching dictates otherwise our data indicate that retrocolic Billroth II anastomoses are preferable to antecolic Billroth II reconstructions after partial gastrectomy for adenocarcinoma of the stomach, as there is a diminished risk of late gastric outlet obstruction and a greater 5-year survival among patients having the former procedure. Survival is unacceptably low after Billroth I anastomoses.