BACKGROUND: Observational studies have suggested that girls have higher mortality if their most recent immunization is an inactivated vaccine rather than a live vaccine. We therefore reanalyzed 5 randomized trials of early measles vaccine (MV) in which it was possible to compare an inactivated vaccines [after medium-titer MV (MTMV) or high-titer MV (HTMV)] and a live standard titer MV (after an initial inactivated vaccine). METHODS: The trials were conducted in Sudan, Senegal, The Gambia and Guinea-Bissau. The intervention group received live MTMV or HTMV from 4 to 5 months and then an inactivated vaccine from 9 to 10 months of age; the control children received inactivated vaccine/placebo from 4 to 5 months and standard titer MV from 9 to 10 months of age. We compared mortality from 9 months until end of study at 3 to 5 years of age for children who received inactivated vaccine (after MTMV or HTMV) and standard titer MV (after inactivated vaccine), respectively. The original datasets were analyzed using a Cox proportional hazards model stratified by trial. RESULTS: The mortality rate ratio (MRR) was 1.38 (95% confidence interval: 1.05-1.83) after an inactivated vaccine (after MTMV or HTMV) compared with a standard titer MV (after inactivated vaccine). Girls had a MRR of 1.89 (1.27-2.80), whereas there was no effect for boys, the sex-differential effect being significant (P = 0.02). Excluding measles cases did not alter these conclusions, the MRR after inactivated vaccines (after MTMV or HTMV) being 1.40 (1.06-1.86) higher overall and 1.92 (1.29-2.86) for girls. Control for variations in national immunization schedules for other vaccines did not modify these results. CONCLUSIONS: After 9 months of age, all children had been immunized against measles, and mortality in girls was higher when they had received inactivated vaccines (after MTMV or HTMV) rather than live standard titer MV (after an inactivated vaccine).