Commentary: Evidence for a 'healthy pregnant woman effect' in Niakhar, Senegal?

E Goodburn
2001 International Journal of Epidemiology  
Since its launch in 1987 the Safe Motherhood Initiative has been successful in drawing attention to the toll of maternal deaths, currently estimated at 600 000 per annum, and their inequitable distribution-99% occur in the developing world. The World Development Report 1993 ranked maternal causes as the leading cause of death and disability among women aged 15-44 in developing countries, accounting for 18% of their disease burden. 1 In certain countries, for example Bangladesh, the paradoxical
more » ... h, the paradoxical excess of female deaths over that of males has been attributed mainly to maternal causes. 2 The general consensus from these findings is that pregnancy in resource-poor settings is a dangerous undertaking. However, despite the growing awareness of maternal deaths as a major public health issue, there remain many uncertainties about the epidemiology of maternal mortality. The many studies that have identified the medical causes of maternal deaths have been remarkably consistent in showing that around 25% are due to indirect causes (disease not due to obstetric causes but aggravated by the pregnancy) while the rest are due to direct obstetric causes. 3,4 Attention has focused on factors contributing to direct obstetric deaths because there are more of them and it is clear that they are largely preventable. We still know very little about the role of indirect causes in maternal deaths and the interactions between disease and pregnancy in high mortality settings. Logic would lead us to expect that death rates from all causes (which would include direct obstetric causes) would be higher in pregnant or recently pregnant women and that, as a number of diseases are known to be aggravated by pregnancy, death rates from disease would also be higher. The findings of the study by Ronsmans et al. which compares the rates and causes of deaths in pregnant and nonpregnant women, challenges many of our assumptions about the risks of pregnancy. 5 In the rural, high-fertility population studied in Senegal, the all-cause death rates in pregnant or recently pregnant women were no higher than in non-pregnant women except at the extremes of the reproductive age group. When direct obstetric deaths were excluded, currently or recently pregnant women were up to five times less likely to die than women who had not been recently pregnant. There is little evidence to suggest that pregnancy itself can confer a protective effect on health status and the authors conclude that their findings are mainly due to a 'healthy pregnant woman' effect. The pregnant population had a greater proportion of healthy women, and, as a result, death rates from non-obstetric causes were lower in this group. This selection bias was important enough in this study to balance out even the excess mortality from causes directly related to the pregnancy. The conclusions of this study may appear surprising, but are not unique to Senegal. A study in rural Bangladesh has produced very similar results. 6 The concept of a 'healthy pregnant woman effect' is strongly supported by research generated by the HIV/AIDS epidemic. There is mounting evidence that women who are severely ill or suffering from chronic diseases are less fertile. 7,8 Although the prevalence of HIV in pregnant women is still very low in Bangladesh and Senegal, these are high mortality populations where other chronic diseases such as TB are common. It could be assumed that, in low fertility, low mortality settings-where many healthy women avoid conception voluntarily by using contraception-that a 'health pregnant woman' selection bias would not be evident. However, results from studies in developed countries show the same effect even though the actual death rates are much lower. 9, 10 The findings of this study serve as a reminder that our understanding of reproductive age female deaths in developing countries, particularly the concept of indirect maternal deaths, is incomplete. One of the reasons for this is that much of our thinking on the epidemiology of maternal deaths has been guided by studies which either have no comparison group or use limited population subsets, e.g. surviving pregnant women, as a comparison group. These studies have been of crucial importance for identifying factors, such as deficiencies in service delivery, which contribute to avoidable deaths. However, the publication by Ronsmans et al. has highlighted the need for further study separating risks attributable to pregnancy from risks to which the whole population is exposed if we wish to explore the role played by pregnancy in the mortality and morbidity of reproductive age women overall. References 1 Starrs A. The Safe Motherhood Action Agenda: Priorities for the Next Decade.
doi:10.1093/ije/30.3.474 fatcat:jyit2jxmsfcrbjwkb4wgqxwggu