Tetanus continues to be a leading cause of neonatal death in Bangladesh as in other developing countries, yet little is known about risk factors or the efficacy of tetanus toxoid in this setting. In May 1990, mothers of 6148 infants born alive between March 15, 1989, and March 14, 1990, in 30 rural unions of Rajshahi Division in Bangladesh were interviewed. Three surviving controls for each neonatal tetanus death were matched for sex, residence, and date of birth. Of 330 neonatal deaths, 112 met the case definition for tetanus. Risk was increased with a history of neonatal tetanus in a previous child, application of coconut oil to the vagina, and use of multiple ties on the umbilical cord. Risk was reduced by the birth attendant washing hands and using a cleaned cord-cutting tool. Risk was not reduced by a maternal history of two doses of tetanus toxoid (TT2), although estimated efficacy of TT2 was 45% (95% confidence interval = 16% to 64%). Subsequent to the survey, a reference laboratory reported to potency in three consecutive lots of tetanus vaccine from the production laboratory in Bangladesh. These findings identify high-risk mothers, stress the importance of washing hands and cleaning the cord-cutting tool, and demand improved quality control of tetanus vaccine production.