Background and design: Osteoporosis has been observed with chronic hypervitaminosis A but has not been established as a toxic effect of synthetic retinoid therapy in humans. This cross-sectional study was designed to assess bone mineral density (BMD) during long-term therapy with the retinoids etretinate or isotretinoin. Twenty-four patients were evaluated for osteoporosis with the standard techniques: single- and dual-photon absorptiometry. They received 50 g or more of etretinate (15 patients) or isotretinoin (nine patients) for 2 years or longer for the treatment of skin disease (ichthyosis [nine patients], Darier's disease [six patients], xeroderma pigmentosum [four patients], skin cancer [three patients], or psoriasis [two patients]). In each of the two treatment groups, BMDs (measured in grams per square centimeter) were measured at five standard sites (ie, lumbar spine, femoral neck, trochanter, Ward's triangle, and radius) and evaluated against a standardized database to control for age, sex, and weight. In addition, for each measurement site, BMDs (controlled for age, sex, and weight) were compared between the two groups, as a direct control for each other.
Observations: Compared with those of the age-, sex-, and weight-matched controls, the BMD values of the etretinate group were significantly decreased at four of the five measurement sites: femoral neck (90.6%, P = .0001), Ward's triangle (87.8%, P = .0001), trochanter (87.8%, P = .0012), and radius (85.0%, P = .039). In contrast, the BMDs in the isotretinoin group did not differ from control values except for an elevation at the lumbar spine (P = .039). When the two groups were compared, the mean BMDs were significantly lower in the etretinate group when measured at the lumbar spine, trochanter, and radius (P < .05).
Conclusions: This study identified osteoporosis in patients who received long-term therapy with etretinate but not isotretinoin. Prospective studies of BMD would be useful to further define retinoid-associated osteoporosis.