Editor—We read with interest the paper by Murphy et al 1 reporting that the decline in prevalence of neural tube defects reached a plateau in the 1990s. In 1991 we reported the results of a study of maternal cohorts in neural tube defects in the Fylde of Lancashire.2 We found that mothers born after 1949 were significantly less likely to have a baby with spina bifida and noted that this was chronologically compatible with fewer births of affected babies after 1973. This fits with the observation that the decline in prevalence seems to have stopped in the 1990s, as by then mothers born before 1950 who were at greater risk of producing a baby with spina bifida, would have passed their reproductive period.

It seems that one cause of the decline in prevalence of neural tube defects between the early 1970s and the early 1990s should be sought in the early 1950s.

We suggested that abandonment of mercury as a therapeutic agent for infants at that time, especially in the widely used teething powders, could explain our findings.