The letters by Drs Whitmore and Margolis criticize the cross-sectional nature of our study,1 which demonstrated an association between etretinate use and decreased BMD. They note that biases or chance occurrence could account for an association when disease assessment and exposure are made at the same time. We agree that pretreatment assessments of BMD would have been desirable, if possible. However, many of these patients have been treated with retinoids since they were first studied in the United States in the 1970s, long before densitometry was commonly available. Radiologic assessment of bone density, a much less sensitive technique, was the standard at the time and is not really a fair comparison.
So then, what are the biases noted and can they realistically explain the association that we identified? Margolis suggests that "it is just as plausible... that the individuals selected for study... had osteoporosis and were chosen to receive
DiGiovanna JJ, Sollitto RB, Steinberg SM, Reynolds JC. Osteoporosis and Long-term Etretinate Therapy-Reply. Arch Dermatol. 1996;132(6):714. doi:10.1001/archderm.1996.03890300144028