Ki-67 immunostaining is commonly used for assessing cell proliferation, but studies of its use as a prognostic indicator have revealed discordant results in gastric cancer patients. Recently, antibodies for phosphorylated histone H3 have been used to identify dividing cells because of its precise overexpression in mitosis. The authors tested the hypothesis that phosphorylated histone H3 overexpression might be a good prognostic indicator for gastric cancer patients by conducting an immunohistochemical comparison with Ki-67 in gastric cancer samples. One hundred twenty-two surgically resected primary cases were selected and histologically categorized in accordance with Lauren's classification. No correlation was found between phosphorylated histone H3 and Ki-67 regarding overexpression. However, correlations between phosphorylated histone H3 overexpression and clinicopathologic variables were noted for histologic type (intestinal type predominant in high labeling indices [LIs], defined as over the value of the 75th percentile; P<0.01), vessel invasion (positive in high LIs; P=0.05), and lymph node metastasis (positive in high LIs; P=0.04). With regard to Ki-67 overexpression, no correlation was evident with the clinicopathologic variables except histologic type (intestinal type predominant; P=0.05). By the Kaplan-Meier method with the log-rank test, cases overexpressing phosphorylated histone H3 showed a poorer prognosis than cases with low expression (P<0.01). In contrast, Ki-67 expression did not influence prognosis. Multivariate analyses indicated phosphorylated histone H3 overexpression to be an independent prognostic factor, together with lymphatic invasion and venous invasion (P<0.01). In conclusion, it seems likely that phosphorylated histone H3 plays an important role in the prognosis of gastric cancer, and its immunohistochemical investigation is useful for the prediction of prognosis in gastric cancer.