Fertility history and cause-specific mortality: a register-based analysis of complete cohorts of Norwegian women and men

Abstract

The relationship between women's reproductive histories and later all-cause mortality has been investigated in several studies, with mixed results. Some studies have also considered cause-specific mortality and some have included men, but none has done both. We analyse associations between parity and age of first birth for women and men across 11 cause-of-death groupings using Norwegian register data for complete cohorts born 1935-1968 whose mortality was observed 1980-2003 (i.e. at ages 45-68). Age, period, educational level, marital status, region of residence and population size of municipality were included as co-variates. In total, there were 63,000 deaths. Results showed that relative to parents of two children, childless men and women and those with one child had higher mortality risks for nearly all cause of death groupings. High parity (4+ children) was associated with raised male mortality from accidents and violence and higher mortality from cancer of the cervix among women. For other cause and gender groupings there was either little difference between those with two children and those of higher parities or an overall negative association between parity and mortality. Among men with the lowest level of education, however, high parity was positively associated with mortality from circulatory diseases. For all causes except female breast cancer, there was an inverse association between age at first birth and mortality risk. Similarities observed across cause groups and for women and men suggest that much of the fertility-mortality relationship is a result of selection or effects of reproductive behaviour on lifestyle. The latter may include both beneficial effects and harmful stress responses. However, physiological mechanisms are most probably important for some causes of death for women. Research on associations between parenting histories, health related behaviours, social support exchanges and reported or measured stress is needed to clarify mechanisms underlying the associations reported here.