Background: Peritoneal, lymph node, and hematogenous recurrence are frequently observed as patterns of recurrence after surgery for gastric cancer. However, the clinicopathological characteristics associated with each recurrence have rarely been comprehensively reported in a multicenter study. Understanding the risk factors for each pattern of recurrence would be helpful for the early detection of recurrence and the initiation of optimal treatment. This study investigated the risk factors for the first recurrence in each pattern after curative gastrectomy, using data from a multicenter randomized controlled trial (JCOG1001) that was designed to investigate the efficacy of bursectomy. Methods: Patients of 20-80 years of age, with cT3(SS)-T4a(SE) gastric carcinoma according to the 14th Japanese Classification of Gastric Carcinoma, with an ECOG PS of 0-1, and a body mass index of < 30 kg/m², and without bulky lymph nodes, Borrmann type 4 or large type 3 carcinoma were eligible for inclusion in JCOG1001. Of the 1204 patients who were enrolled in JCOG1001, 932 pStage II/III patients with a common histological type were included in this study. Risk factors for hematogenous, lymph node, and peritoneal patterns of recurrence were estimated by a multivariable Fine and Grey model considering death or site of recurrence other than the first site of recurrence as competing risks. Results: The overall rate of recurrence was 27.1%. Hematogenous recurrence was the most frequent pattern (12.3%), followed by peritoneal (11.2%) and lymph node (7.5%) recurrence. Differentiated type (HR, 1.818; 1.237-2.674; p = 0.0024), pT4 (in comparison to pT1-3, HR, 1.511; 95% CI, 1.011-2.257; p = 0.0440), and pN3 (in comparison to pN0-2, HR, 2.431; 95% CI, 1.635-3.616; p < 0.0001) were associated with an increased incidence of hematogenous recurrence. Conversely, more than D2 lymphadenectomy reduced this pattern of recurrence (in comparison to D1+or D2 lymphadenectomy, HR, 0.575; 95% CI, 0.364-0.907; p = 0.0174). Peritoneal recurrence was significantly associated with large (≥5 cm) tumor (HR, 1.649; 95% CI, 1.034-2.629; p = 0.0356), pT4 (in comparison to pT1-3, HR, 3.222; 95% CI, 2.086-4.976; p < 0.0001), pN3 (in comparison to pN0-2, HR, 1.865; 95% CI, 1.275-2.727; p = 0.0013), and undifferentiated type (HR, 2.674; 95% CI, 1.628-4.394; p = 0.0001). Extended lymph node metastasis (pN3) was the only risk factor (in comparison to pN0-2, HR, 8.030; 95% CI, 4.605-14.002; p < 0.0001) for lymph node recurrence. Conclusions: The risk factors for recurrence differed according to the patterns of recurrence. Vigilant follow-up with an understanding of patterns of recurrence is required, especially for high-risk patients.