The long-term effects of child bearing on adult mortality in Matlab, Bangladesh

It is argued that reproduction impairs female survival in the long term, as pregnancy is thought to drain maternal physical resources. However, a review of the literature revealed that there was little evidence to suggest that mortality of highly parous women exceeded that of less parous women. In fact, mortality appeared to be highest in nulliparous women or women with only one birth. In addition, women whose first births were later were found to have consistently lower mortality. There was no evidencet hat female reproductive histories affected the risk of mortality in males. The aim of this thesis was to examine whether reproductive history was associated with all-cause mortality after age 45 in women and men who had completed their reproduction in Matlab, Bangladesh. A cohort study was conducted using demographic data collected by the International Centre for Diarrhoeal Disease Research, Bangladesh. Essential data were found to be missing in the sample and multiple imputation was used in an attempt to adjust for the potential bias that this missing data may have introduced. Trends in crude mortality rates with the reproductive exposures were examined. Poisson regression was then used to examine the association between reproductive history and all-cause mortality, adjustingf or potential confounders Crude mortality rates for the reproductive variables, stratified by age, education, religion and marital status, were also examined for effect modification and likelihood ratio tests for interaction performed. There were no differences in the mortality of women with parity, either when comparing the parous with the nulliparous or when looking at trends with the number of children born. However, mortality did decrease with the number of surviving children. These differences were statistically significant and persisted when the rate ratios were adjusted for potentialc onfounders. The patterns in the husbands were strikingly similar to those seen in the women. The results suggest that there are few negative long-term biological consequences of bearing children in these women. Even if they exist, they may be outweighed by the social advantages of having a fiunily and a healthy pregnant woman effect. These conclusions are strengthened by the fact that the reduction in mortality with the number of surviving children persisted after adjusting for potential confounders and by the remarkable consistency in the associations in the females and their husbands.