Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
We were interested in a recently published article in Archives by Conrad and colleagues.1 They concluded that infants fed soy formula had a prolonged increase of thyroid stimulating hormone when compared to infants fed by non-soy formula. We have some criticisms of their study methods.
In this retrospective study there was a notable difference between the patient numbers in the soy diet group (n = 8) and non-soy diet group (n = 70). It is well known that in prospective studies in which data of two groups are compared, in order to gain statistically significant results there should be a minimum of 10 test subjects in each group and the numbers in the groups should be close. Although it is not essential to follow this rule in retrospective studies like the one of Conrad et al, the statistical reliance of the study fails since the soy diet group has eight patients whereas the other one has 70.
Secondly, in studies in which comparisons of any of body fluid parameters are made for each group, for better results, it is important that the materials must be studied in the same sessions using calibrated machines after the materials have been stored appropriately. This could not be achieved since the study was retrospective, and it is therefore inevitable that there were differences between the thyroid stimulating hormone and thyroxine results of the soy diet and non-soy diet groups. For these two reasons we think it is impossible to conclude that soy formula decreases the success of treatment in congenital hypothyroidism. We believe that further prospective controlled studies can better shed light on this topic.