For more than 15 years, Peter Aaby and colleagues have been studying mortality after infant vaccination in various sub-Saharan African populations and describing ‘non-specific’ or ‘heterologous’ effects of the vaccines on infant death.1,–,11 In summary, Aaby's observations suggest that live vaccines such as BCG and measles reduce all-cause mortality (ie, mortality caused by afflictions other than tuberculosis or measles infection) in the period after they are administered, while, at least in girls, diphtheria-tetanus-pertussis (DTP) containing vaccines increase mortality.

Before discussing this phenomenon further, it must be taken in the context of a remarkably successful WHO expanded programme on immunisation, which currently prevents more than 2 million childhood deaths every year12 and will do more in the decade ahead as vaccines continue to be rolled out against Haemophilus influenza type b, rotavirus and pneumococcus. Indeed, one of the main obstacles to open discussion of these non-specific vaccine effects, and independent investigation of the issue, is that an adequate account of the positive impact of vaccines has not been included and, as a result, the phenomenon is perceived as a threat to the life-saving global immunisation vision and strategy with the result that the topic has been largely avoided by funders and investigators.

From an immunological perspective, the hypothesis that there could be non-specific effects of vaccines is not so controversial, and there are several lines of evidence that have shown that variations in the immune response …