First,case control studies using selfreported data on vaginal bleeding appear to suffer from underreporting. Moreover ,data on cigarette smoking were based on recall at the time of the time of delivery in most Of these case-control studies, thereby introducing a potential for recall bias.