The correspondents mention some of the most important harmful effects that have been associated with multiple courses of antenatal steroids, and a growing number of studies have demonstrated such effects. Despite this, it should not be forgotten that multiple courses of antenatal steroids may have beneficial effects, and the assumption underlying their widespread use may be correct1,2. It is possible that both beneficial and harmful effects could be caused simultaneously3. A greater number of recent studies have suggested harm than have suggested benefit, but this could be because more studies have investigated adverse outcomes, or that there has been selective reporting of results suggesting harm. Until randomised trials of single versus multiple courses are carried out, the balance of risks and benefits of this therapy will remain unknown.
Existing evidence of the effects of multiple courses on neonatal and maternal outcomes derives from two sources: animal experiments and observational studies of humans. The correspondents mention some of the possible problems with animal studies; in addition it is unknown whether any of the structural changes that have been found in animal brains exposed to multiple courses of steroids occur in humans, nor whether this results in neurodevelopmental problems. The applicability of these results to human is therefore unknown.
Observational studies of humans are also problematic, as they are open to potentially serious biases. First, the number of courses that an infant is exposed to is confounded with the gestational age at birth. Infants exposed to more courses will tend to be born at greater gestational ages and have better outcomes for this reason. Second, infants exposed to multiple courses may remain in a high-risk situation for longer and hence have poorer outcomes than those exposed to only one course. Depending on which source of bias is stronger in any particular study, the results may be influenced in either direction, either to favour single courses or to favour multiple courses, and indeed observational studies have produced conflicting results. Therefore, observational studies cannot give an unambiguous estimate of the risks and benefits of multiple courses of antenatal steroids, and randomised trials are necessary.