A double-blind, randomized, Phase III, multicenter study in 358 pediatric subjects receiving isotretinoin therapy demonstrates no effect on pediatric bone mineral density

The Summary This study compared the effects of pediatric acne treatment with two isotretinoin formulations on bone mineral density. We demonstrated no difference in the effect of the two formulations. No effect on pediatric bone mineral density was identified for either formulation. Introduction Isotretinoin (13-cis-retinoic acid) is a treatment for recalcitrant nodular acne with a purported effect on bone mineral density (BMD). The side effects of isotretinoin on vertebral bone were evaluated to assess the safety of a new FDA-approved isotretinoin formulation: Lidose-isotretinoin (Cip-Iso). Methods This double-blind, randomized, phase III, active control, parallel-group, multicenter study compared the safety, efficacy, and non-inferiority of CIP-Iso to a marketed reference product, AccutaneA (R), in severe recalcitrant nodular acne subjects. Three hundred fifty-eight pediatric male and female subjects aged between 12 and 17 years underwent 20 weeks of treatment with PA lumbar spine dual X-ray absorptiometry (DXA) measurements obtained for bone mineral density (BMD) and Z-scores, 5.5 months apart on visits 1 and 8. One hundred sixty-eight of 358 subjects had height adjusted Z-scores (HAZ) calculated. Results There was no difference in the least squares (LS) mean Z-score or HAZ of the two drugs at visit 1 or 8. The mean and LS mean Z-score and HAZ were greater than zero at visits 1 and 8 for both drugs. The change in the LS mean spine Z-score, but not HAZ, between visits, was statistically significant for both drugs. There was a mean increase in BMD (g/cm(2)) for both products between visits. Conclusions There is no difference in the effect of two formulations of isotretinoin on spine bone density after 6 months of treatment. BMD increased and the small change in spine Z-score over treatment disappeared after height adjustment. Mean positive Z-scores and HAZ in the study were likely due to the exclusion of low and inclusion of high Z-score subjects.