Today on msn.com I found metion of this article reporting a study of the effect of breastfeeding on the stress response of children at 10 years of age. The study involved almost 9000 children born in Great Britain in 1970. The occurrence of divorce or separation of the parents was also included in the statistical analysis, along with a few other variables such as maternal age and the sex of the child. As predicted, children whose parents had divorced or separated between when the child was 5 and 10 years old had higher levels of anxiety. In addition, children who had been breast-fed had lower anxiety than those who had been bottle-fed. Interestingly, the effect of parental divorce or separation on anxiety was only significant for the bottle-fed children. Parental separation and divorce were not correlated with breastfeeding themselves, and none of the other variables appeared to mediate the effect of breastfeeding.

I likely wouldn't have thought anything of this study before, but after reading about methylation of the glucocorticoid receptor gene in rats being affected by maternal behavior (see the "alter their response to stress" link in this old GNXP post), this makes me wonder if a similar effect operates in humans. Although there are many ways to explain the human child study that don't invoke this type of biochemical process, the pattern of altered responsiveness to an environmental event (parental separation/divorce) would be consistent with some kind of alteration in the stress response pathway itself.