Childbearing history and late-life mortality: the Dubbo study of Australian elderly

Objective: to examine the association of parity with mortality in later life. Design: a longitudinal, community-based study. Setting: semi-rural town of Dubbo, NSW, Australia. Subjects: a total of 1,571 women and 1,233 men 60 years and older first examined in 1988–89. Outcome measures: all-cause and cause-specific mortality rates analysed over 16-year follow-up. Hazard ratios obtained from proportional hazards models employing conventional predictors, potential confounders and measure of parity. Results: increasing parity in women was weakly associated with overweight, diabetes and hypertension. All-cause mortality fell progressively with increasing parity in women (hazard ratio and 95 % confidence intervals): childless, 1.00; 1 child, 1.03 (0.75–1.43); 2 children, 0.83 (0.61–1.11); 3 children, 0.80 (0.60–1.08); 4 children, 0.91 (0.66–1.25); 5 children, 0.70 (0.49– 1.01); 6+ children, 0.60 (0.43–0.85) (trend for parity P < 0.002). This result was similar whether or not hypertension, dia-betes and overweight were included in multivariate models adjusting for social variables and other confounders. The reduc-tion in all-cause mortality was accompanied by a parallel reduction in deaths from cancer and respiratory conditions, while coronary heart disease mortality increased 60–111 % in all parous women. Conclusion: there was increased all-cause mortality in later life in childless women, accompanied by reduced mortality as parity increased. Underlying mechanisms are unclear but findings may have public health importance