Historically, the factors associated with declines in infant and early childhood mortality in the nineteenth and early twentieth centuries seem to have varied from country to country. Around the beginning of the twentieth century, for instance, child mortality was high in the USA and, in contrast to the situation in the UK, the rich and advantaged did not fare much better than the poor. In England and Wales, infant mortality was greater in urban areas, a fact attributed in the main to diarrhoea and lack of adequate sanitation. A study in the Netherlands, however, (International Journal of Epidemiology2000;29:1031–40) has shown no “urban-sanitary-diarrhoeal-effect” on infant mortality in the last quarter of the nineteenth century. Urbanisation there had the greatest impact on post-infantile, early childhood mortality from acute respiratory diseases. These authors stress the association between religion and infant mortality rates in the Netherlands at that time. The rates were higher in Roman Catholic areas and this they attribute to lower rates of breast feeding, adherence to folk medicine, and reluctance to accept newer ideas about medicine and public health.

Data from a 1997 study in Pakistan (American Journal of Clinical Nutrition 2000;72:1164–9) and from the third US National Health and Nutrition Examination Survey (NHANES III) of 1988–94 (Ibid: 1170–8) confirm that infection lowers serum concentrations of retinol and knowledge of serum concentrations of acute phase proteins …