In a double-blind study 10 patients with oral leukoplakia were treated daily (three topical applications) with 0.1% isotretinoin gel or a placebo for 4 months. Nine patients completed the treatment, while one patient was lost to follow-up. Subsequently, the patients who had received the placebo used the active medication for an additional 4 months. All patients treated with the active medication showed a significant improvement of the oral lesions while, in the patients receiving only the placebo, the size of the lesions remained the same. Also the group of patients who received the active medication after the placebo showed an improvement in the size and clinical appearance of the lesions. A complete response was defined as the complete disappearance of the lesion as assessed by visual inspection, while a partial response was defined as a 50% or more reduction in the size of the lesions. In total we had a complete response and eight partial responses. No side-effects from the use of the gel were ever observed. The percentage of bcl-2-positive cells was evaluated in the basal layer from a minimum of 1000 cells in each case and the bcl-2 immunostaining was scored using three groups: - (< or = 10% cells); + (< or = 50% cells); +2 (> or = 50% cells). The presence of apoptotic bodies was evaluated in a random fashion in the parabasal layer in 20 HPF. Immunohistochemical analysis for bcl-2 protein showed that before treatment a weak positivity of the basal layers, with a focal positivity of some parabasal cells, was present: five out of nine specimens were positive. Only a few apoptotic bodies were observed. After treatment in almost all specimens it was possible to observe a complete bcl-2 negativity with a positivity in only one specimen out of nine. An increase in apoptotic bodies was observed. Statistical analysis showed that the difference between the bcl-2 positivity in the two groups was not statistically significant (P = 0.134) while, on the contrary, the difference in the count of the apoptotic bodies between the same two groups was statistically significant (P = 0.0193). In conclusion, the data obtained from this pilot study show that good results can be obtained with the topical use of 13-cis-retinoic acid.