RESUMO
OBJECTIVE: To determine whether vaccination against measles in a population with sustained high vaccination coverage and relatively low child mortality reduces overall child mortality. METHODS: In April and May 2000, a population-based, case-control study was conducted at Ballabgarh (an area in rural northern India). Eligible cases were 330 children born between 1 January 1991 and 31 December 1998 who died aged 12-59 months. A programme was used to match 320 controls for age, sex, family size, and area of residence from a birth cohort of 15 578 born during the same time period. FINDINGS: The analysis used 318 matched pairs and suggested that children aged 12-59 months who did not receive measles vaccination in infancy were three times more likely to die than those vaccinated against measles. Children from lower caste households who were not vaccinated in infancy had the highest risk of mortality (odds ratio, 8.9). A 27% increase in child mortality was attributable to failure to vaccinate against measles in the study population. CONCLUSION: Measles vaccine seems to have a non-specific reducing effect on overall child mortality in this population. If true, children in lower castes may reap the greatest gains in survival. The findings should be interpreted with caution because the nutritional status of the children was not recorded and may be a residual confounder. "All-cause mortality" is a potentially useful epidemiological endpoint for future vaccine trials.