BACKGROUND: Due to delays in vaccinations, diphtheria-tetanus-whole-cell-pertussis (DTP) is often given with or after measles vaccine (MV) - out of sequence. We reanalysed data from Matlab, Bangladesh, to examine how administration of MV and DTP out-of-sequence was associated with child survival.

METHODS: 36,650 children born between 1986 and 1999 were followed with registration of vaccinations and survival. Controlling for background factors using Cox proportional hazards models, survival was analysed between 9 and 24 months of age. We measured the mortality rate ratio (MRR) to compare vaccination groups. Oral polio vaccine (OPV) campaigns, which started in 1995, reduced the mortality rate and reduced the difference between vaccination groups. In the main analysis, we therefore censored for OPV campaigns; there were 151 non-accidents deaths before the OPV campaigns.

RESULTS: Compared with MV administered alone (MV-only), DTP administered with or after MV had MRR 2.20 (1.31-3.70), and DTP-only had MRR 1.78 (1.01-3.11). Compared with MV-only, DTP administered with MV had a female-male MRR 0.56 (0.13-2.38), significantly different to DTP administered after MV which had MRR 14.83 (1.88-117.1), test of interaction p=0.011. Compared with having DTP (no MV) as most recent vaccination, MV-only had a non-accident MRR of 0.56 (0.32-0.99).

CONCLUSION: The negative effects of non-live DTP with or after live MV are not explained merely by selection bias. These observations support a live-vaccine-last policy where DTP should not be given with or after MV.