Though previous studies have suggested a non-specific beneficial effect of oral polio vaccine (OPV), there has been no evaluation of the mortality impact of national polio immunization days. On the other hand, studies examining the effect of OPV and diphtheria-tetanus-pertussis (DTP) vaccines, which are usually administered together in routine immunisation programmes in low-income countries, have found no beneficial or even a negative effect on infant survival. In 1998, we used the opportunity of two national immunisation days to examine the impact of OPV administered alone on survival for the 6103 children less than 5 years of age in the Bandim Health Project's study area in Guinea-Bissau. Survival was ascertained through regular surveillance from March 1998 until the beginning of the war on June 7, 1998, the end of 1998, or the end of 1999, respectively. The child register was linked with a register for the only paediatric ward in Bissau to determine the risk of hospitalisations. Among children under 5 years of age, 82% had received 1 or 2 doses of polio vaccines during the campaign. Though polio vaccination during the campaign was associated with slightly lower mortality, this difference was not significant for all children under 5 years of age (mortality ratio (MR)=0.46 (0.18-1.15)). However, oral polio vaccination was associated with a beneficial effect for children under 6 months of age at the time of the campaign, the mortality ratio being 0.09 (95% CI 0.01-0.85) in the 3 months before the war controlling for significant background factors, including routine immunizations, antenatal consultations, and arm circumference. The polio-vaccinated children aged 0-5 months had fewer hospitalisations than children who had not been polio vaccinated (RR=0.27 (0.10-0.76)). With longer follow-up to December 1998 or December 1999, the difference in mortality gradually disappeared, the MR for polio-vaccinated children being 0.61 (0.32-1.14) and 0.83 (0.51-1.34), respectively. Among children aged 6-59 months of age, measles vaccine was associated with a 56% reduction in mortality (MR=0.44 (0.28-0.69)) and no effect of oral polio vaccine was measurable in this age group. The effect of polio vaccine among children less than 6 months of age could be due to selection bias but might also represent a non-specific beneficial immune stimulation and there is nothing to suggest that OPV might have a negative effect on infant survival. Studies of the possible non-specific effects of oral polio vaccine are warranted before OPV is withdrawn.