PURPOSE: We compared the safety of the double-stapling method with that of the hand-sewn method when performing a Billroth I anastomosis (BI) after a radical subtotal gastrectomy. METHODS: Fifty-three (53) patients who received a BI with stapler from February 1998 to April 1999 were compared with 24 patients who received a BI with the hand-sewn method from January 1996 to December 1998. All the patients were divided into three groups according to the type of reconstruction: Group 1 (n=42), double- stapling method, Group 2 (n=11), anastomosis through posterior gastric wall, and Group 3 (n=24), hand- sewn method. The length of the greater curvature, lesser curvature, and proximal margin and the number of resected lymph nodes were compared for radicality, and the incidence of complications was reviewed. ANOVA and the chi-square test were used for the statistical analysis. RESULTS: The mean ages were 55.2+/-10.8, 59.8+/-9.7, and 54.7+/-11.9 years for groups 1, 2, and 3, respectively. The sex ratios were 2.2:1, 1.2:1, and 1.4:1. The lengths (cm) of the greater curvature were 17.2+/-2.5, 18.4+/-2.5, and 18.6+/-3.3, and the lengths of the lesser curvature were 10.4+/-1.5, 10.9+/-1.6, and 10.7+/-2.1; however, the differences were not statistically significant (p>0.05). The lengths of the proximal resection margin were 5.4+/-2.5, 8.2+/-3.3, and 6.0+/-3.2 (p<0.05), and the numbers of resected lymph nodes were 25.3+/-8.3, 19.1+/-5.8, and 32.+/- 13.7 (p<0.01). We experienced no complication in any group, except one case of leakage in the hand-sewn group. CONCLUSION: The double-stapling technique is a safe operative method in BI after a radical subtotal gastrectomy with its very low complication rate. Especially, the so- called Tornado method is a more useful method since it doesn't need a gastrotomy in the remnant stomach.