Since November 1995 we have been performing a D3 lymph node dissection in patients undergoing an operation for gastric cancer with a curative intent. The aim of the present study was to evaluate whether this procedure results in an increased postoperative mortality or complication rate in a Western population. Between November 1995 and August 1997 the postoperative courses of 76 patients were retrospectively assessed (45.3 lymph nodes per patient, lymph node ratio: 0.16). The patient outcome was compared with data from a historic control group of patients (n = 383) in whom the newly established D2 dissection was studied in our department. Regarding the demographic, clinical, and tumor-pathologic data, and the choice of resection and reconstructive procedures, the two groups differed only slightly. The postoperative mortality of 1% was lower (vs 6.8%) while the overall complication rate of 34% (vs 32.1%) was identical. In particular, no anastomotic leakage (vs 9.4%) and fewer nonsurgical complications (17.1% vs 27.9%) occurred. The reoperation rate was 1% vs 9.7%. However, in 6% of the patients drainage tubes had to be inserted under computed tomographic guidance. The average hospital stay remained unchanged (21.9 vs 20.7 days). A D3 dissection was shown to be feasible while demonstrating no disadvantages in the patients when compared with the D2 procedure.