Previous reports on the influence of the sex of a child and prognosis of a subsequent breast cancer have been conflicting. We took advantage of a number of large and good quality, nationwide Registries in Sweden to evaluate the prognostic value of the sex of the first child in breast cancer. A population-based cohort of 32,003 women born after 1935 and diagnosed with primary breast cancer between 1958 and 1997 was generated by linking a number of Swedish registries including Swedish Cancer Registry, Cause of Death Registry, Swedish Generation Registry and the Registry of Population and Population changes. Of these, 4,251 were nulliparous, 2,732 had only a boy and 2,497 had only a girl prior to diagnosis. We used this subpopulation to quantify the association between the sex of the first child and breast cancer-specific mortality, using the Cox proportional hazards. Sex of the child did not significantly influence the prognosis of breast cancer. Stratifying by age and adjusting for age at diagnosis, calendar period, time since last birth and age at first child did not alter the results. Our study, the largest up to date, failed to identify evidence to support the proposed association between the sex of the first child and breast cancer survival. Given the previous reports that advocated taking sex of the first child into consideration in making decisions related to breast cancer therapy, our findings are reassuring and clinically important. ' 2005 Wiley-Liss, Inc.Key words: gender of first child; breast cancer survival; prognosis; epidemiologyThe sex of the child should significantly affect the prognosis of a subsequent breast cancer was first reported by Juret et al.1 in a Lancet communication in 1978. Women whose first child was male had a better survival than those whose first child was female. Contradicting the findings of Juret et al.,1 two studies have reported better survival in women whose first child was female rather than male.2,3 Another study found no association, 4 while a 4th study reconfirmed the first report.5 Thus, the issue of whether the sex of the child influence the prognosis of breast cancer remains unsettled. Apart from varying strategies for selection of study population, insufficient sample sizes and inaccurate ascertainment of the birth history might explain these contradictory findings.We linked nationwide registries including Swedish Cancer Registry and Multi-Generation Registry, which gave us access to all incident breast cancers with complete follow up and objective ascertainment of sex, number and dates of all births, to evaluate the prognostic value of the sex of the first child in women with breast cancer, stratified by age at diagnosis.
MethodsSince January, 1, 1947, the Civil Registration System in Sweden has assigned an individually unique registration number to all residents and citizens. This number is used in all national registries and permits accurate linkage of information obtained from different registries in Sweden. Our analyses are based on a database that was created...