BACKGROUND: The need for lymph node dissection in early gastric cancer (EGC) is controversial. The present study investigated the possibility of planning treatment for EGC according to age and sex rather than node status. METHODS: Overall survival rate and cause of death were analysed according to age (5-year increments) and sex in 4231 patients with EGC. Cox proportional hazard regression analyses were used to identify the most valuable predictor. RESULTS: In patients with EGC 5- and 10-year cancer-specific survival rates were 98.4 and 96.3 per cent respectively, whereas corresponding overall survival rates were 90.2 and 80.9 per cent. The critical age for determining prognosis was 70 years for men (chi2 = 131.34, P < 0.001) and 75 years for women (chi2 = 64.35, P < 0.001). For both sexes, the 10-year overall survival rate was less than 30 per cent in patients over 80 years old. Multivariate Cox stepwise regression analysis identified age as the most powerful prognostic indicator in EGC. The rate of death from causes unrelated to the tumour increased significantly with age, whereas that from recurrence was not affected by age. CONCLUSION: Age is a better prognostic indicator than node status in both men and women with EGC. Age and sex should be taken into account as well as conventional clinicopathological variables related to lymph node metastases when determining appropriate therapy for EGC. Copyright 2004 British Journal of Surgery Society Ltd.
BACKGROUND: The need for lymph node dissection in early gastric cancer (EGC) is controversial. The present study investigated the possibility of planning treatment for EGC according to age and sex rather than node status. METHODS: Overall survival rate and cause of death were analysed according to age (5-year increments) and sex in 4231 patients with EGC. Cox proportional hazard regression analyses were used to identify the most valuable predictor. RESULTS: In patients with EGC 5- and 10-year cancer-specific survival rates were 98.4 and 96.3 per cent respectively, whereas corresponding overall survival rates were 90.2 and 80.9 per cent. The critical age for determining prognosis was 70 years for men (chi2 = 131.34, P < 0.001) and 75 years for women (chi2 = 64.35, P < 0.001). For both sexes, the 10-year overall survival rate was less than 30 per cent in patients over 80 years old. Multivariate Cox stepwise regression analysis identified age as the most powerful prognostic indicator in EGC. The rate of death from causes unrelated to the tumour increased significantly with age, whereas that from recurrence was not affected by age. CONCLUSION: Age is a better prognostic indicator than node status in both men and women with EGC. Age and sex should be taken into account as well as conventional clinicopathological variables related to lymph node metastases when determining appropriate therapy for EGC. Copyright 2004 British Journal of Surgery Society Ltd.