Non-targeted effect of BCG: Several recent studies suggest that BCG has beneficial non-targeted effects on general child survival in low-income countries. Studies of the effect of BCG on morbidity in humans are scarce; some found a positive effect of BCG and others show no effect. Non-targeted effects of vaccines-possible bias and confounding: The major argument against comparing vaccinated and unvaccinated groups is that there is a beneficial social selection bias for vaccinated children-the "Healthy vaccinee effect". However, controlling for various social and health-related background factors in the survival analyses had no effect on the estimates, making this source of bias less likely. A more powerful argument that the findings are not due to the healthy vaccinee effect is that differential non-targeted effects of other vaccines have been observed; diphteria-tetanus-pertussiss vaccination has marked negative effects on child survival, whereas measles vaccine has a positive effect in several studies. Several studies have shown better survival for children reacting to their BCG vaccination with a BCG scar or tuberculin skin test reaction (TST). It could be argued that the reacting children were immunologically stronger and therefore more likely to survive-the "Healthy reactor effect". However, recent findings show that a BCG scar and a TST reaction depend to a large extent on the vaccination technique. Hence, the BCG responses may reflect a true vaccine effect and not merely the health status of the children. Since HIV-1 has been shown to suppress both TST and BCG scar reaction in response to BCG, it is an obvious contributor to the healthy reactor effect, but excluding deaths of children with HIV-1 infection from analysis did not affect the beneficial effect of having a positive TST. Excluding children exposed to tuberculosis (TB) in the household did not affect the estimates either. Furthermore, there are strong sex-differential effects of BCG in both mortality and morbidity data, BCG being more beneficial for girls. These observations cannot consistently be explained by the healthy vaccinee or healthy reactor effects. Ethical implications: For future TB-vaccine studies, these findings imply that: These recommendations might be considered to delay or to be a too large obstacle for the development and trials of new TB vaccines. However, most of the non-targeted beneficial effects of BCG have been observed in children below 2 years of age, which is not a long follow-up time in a TB-vaccine trial. Furthermore, considering the difficulty in setting the TB diagnose in children and the lack of reliable TB-protection markers, it does not seem unreasonable to argue for monitoring of general morbidity and survival in future TB-vaccine trials.