Editor,—Fleming and colleagues state the protective association between pacifiers and cot death is probably real, but feel recommendations on its use cannot be made.1 Is the association causal? There are several hypotheses that might provide a biological mechanism, and thus strengthen this supposition. Having a pacifier might prevent turning prone face straight down. Or perhaps pacifiers facilitate switching to mouth breathing if nasal occlusion occurs.2 Both mechanisms would explain why only pacifier use on the last occasion is protective. Usually using a pacifier would then only be significant if highly correlated with use in reference sleep. A dose–response effect of pacifiers could only be expected if the underlying mechanism would need repeated use to be effective—for example, through repetitive sucking which would increase muscle tone and thus oropharyngeal patency. A dichotomy between never or only rarely using a pacifier and using it always or often would then be more logical than the study's dichotomy ever/never.

Studies on the risks and benefits of pacifiers are hampered by the issue of reverse causality. Do pacifiers increase the risk of otitis media? Or do mothers try to soothe their infant with a pacifier when it suffers from (recurrent) otitis? Does pacifier use have an adverse effect on breast feeding? Or is it a marker for breast feeding difficulties or an attempt to wean the baby? A definitive answer can only be given by randomised trials where pacifiers are introduced at a set time, but clearly these are not easy to carry out. It may not be possible to postpone a decision on pacifiers until such trials are conducted, if indeed they ever will be. Surely the major potential disadvantage of pacifier use is its effect on breast feeding. This needs to be explored further. With current knowledge we would think, however, that using a pacifier can be recommended actively for infants that are bottle fed only.