BACKGROUND Previous case-control studies of neonatal tetanus (NNT) in the North West Frontier Province of Pakistan indicated that clarified butter (ghee) applied to the umbilical wound of newborns was a significant risk factor for NNT. However, the mechanisms underlying the risk remained undisclosed. METHODS A hospital-based case-control study was undertaken to evaluate further ghee and other factors possibly associated with risk of NNT. Mothers of several recent ghee-associated cases were visited in their homes, asked to simulate the procedures used in preparing the ghee, and samples of ghee were collected for culture. RESULTS Topical application of ghee to the umbilical wound was again shown to pose a significant risk for NNT. In-use contamination of ghee was documented as mothers repeatedly heated and manipulated samples of ghee set aside in special containers for this purpose. Ghee was usually applied to the umbilical wound of the baby several times each day for the first few days of life. Mothers of cases were again confirmed to be substantially more likely to report prior NNT cases than mothers of controls. CONCLUSIONS Educational interventions to reduce umbilical ghee use or to wash hands before each manipulation might reduce the risk of NNT in babies exposed to ghee who are born to non-immunized mothers. Increased efforts to immunize women of childbearing age with tetanus toxoid are also needed, with special priority for mothers known to have been associated with a previous NNT case. Topical antibiotics should be further evaluated for protective effects in non-immunized mothers.