Background: Randomized trials have shown that measles vaccine (MV) prevents nonmeasles deaths. MV campaigns are conducted to eliminate measles infection. The overall mortality effect of MV campaigns has not been studied.
Methods: Bandim Health Project (BHP) surveys children aged 0-4 years in rural Guinea-Bissau through a health and demographic surveillance system. A national MV campaign in 2006 targeted children aged 6 months to 15 years. In a Cox proportional hazards model with age as the underlying timescale, we compared mortality of children aged 6-59 months after the campaign with mortality in the same age group during the 2 previous years.
Results: Eight thousand one hundred fifty eight children aged 6-59 months were under BHP surveillance during the 2006 campaign and 7999 and 8108 during similar periods in 2004 and 2005. At least 90% of the eligible children received MV in the campaign. There were 161 nonaccident deaths in 12 months after the campaign compared with 203 and 206 deaths in the 2 previous years, the adjusted mortality rate ratio (aMRR) comparing all children in 2006 with all children in 2004 to 2005 being 0.80 (95% confidence interval: 0.66-0.96). Censoring deaths caused by measles infection, the aMRR was 0.83 (0.69-1.00). The mortality reduction was separately significant for girls [aMRR = 0.74 (0.56-0.97)] and for children who also had received routine MV [MRR = 0.59 (0.36-0.99)].
Conclusions: Mortality levels were stable during 2004 and 2005, but a significant drop occurred after the 2006 MV campaign and was not explained by the prevention of measles deaths. If MV campaigns reduce nonmeasles-related mortality, the policies for measles vaccination should take this into account.