Sex as a prognostic factor in gastric cancer

The aim of this study was to assess whether survival of gastric cancer patients differed between males and females. Although it is well known that the incidence of gastric cancer is higher for men than for women, the existence of a sex-specific prognosis has seldom been addressed. Studies based on population registries have not assessed the role of stage and histology. Cases of histologically confirmed gastric carcinoma were obtained from three Spanish hospitals in Soria (n = 405), Barcelona (n = 249) and Mataro (n = 197). Differences in possible confounders were tested between men and women and survival analyses were performed separately by hospital. Cox's proportional hazards models were used to account for age, tumour stage, histology and tumour sub-location. Only in Mataro was a significant difference in the stage distribution observed between women and men, with a lower proportion of local stage tumours among women (P = 0.047). No statistically significant differences of histological type between men and women were observed in any of the centres. After adjusting for tumour stage and age, women were observed to have significantly better survival in Barcelona (female to male hazard ratio (HR) = 0.578, P < 0.001); this effect was marginal in Soria (HR = 0.788, P = 0.092) and nonsignificant in Mataro (HR = 0.895, P = 0.54). Age-adjusted hazard ratios were calculated within each tumour stage. For Barcelona, the effect of better prognosis among women was most marked at local stage (HR = 0.320, P = 0.013), and in Soria at the regional stage (HR = 0.426, P = 0.002). Although in Mataro all HRs were below unity, none were statistically significant. Little effect was observed at the disseminated stage. The other covariables exerted no influence. Women appear to have a better prognosis than men, and the difference could be tumour stage dependent. Confirmation of these findings would give a valuable insight into gastric cancer growth and ultimately be of use in planning treatment.