In the current study, rats were made diabetic with streptozotocin (STZ) and maintained for 8 weeks, during which time they were treated topically on alternative days with a solution of 0.1% all-trans retinoic acid in a vehicle of 70:30% ethanol/propylene glycol. STZ-induced diabetic rats treated with vehicle served as controls. Additional nondiabetic rats were treated with all-trans retinoic acid or vehicle in parallel. At the end of the 8-week period, rats from all four treatment groups were subjected to abrasion wound formation. Wounds healed more rapidly in vehicle-treated nondiabetic skin than in vehicle-treated diabetic skin (96% of the wound surface area closed in nondiabetic rats within 6 days vs. 41% closed in diabetic rats). Wounds in all-trans retinoic acid-treated diabetic skin healed more rapidly than wounds in vehicle-treated diabetic skin (85% of the wound surface area closed in all-trans retinoic acid-treated diabetic rats vs. 41% closed in vehicle-treated diabetic rats). At the histological level, recently healed skin from vehicle-treated diabetic rats was shown to contain a thin, wispy provisional matrix in which many of the embedded cells were rounded and some were pycnotic. In contrast, a much denser provisional matrix with large numbers of embedded spindle-shaped cells was observed in healed wounds from diabetic skin that had been pretreated with all-trans retinoic acid. The all-trans retinoic acid-treated diabetic skin was histologically similar to vehicle-treated (or all-trans retinoic acid-treated) skin from nondiabetic animals. In light of these findings, we suggest that prophylactic use of retinoid-containing preparations might be useful in preventing the development of nonhealing skin ulcers resultant from minor traumas in at-risk skin.