In a prospective, double-masked clinical study, the authors evaluated the efficacy and safety of topically applied tretinoin ophthalmic ointment (0.01%) versus placebo in the treatment of squamous metaplasia associated with various ocular surface disorders involving dry eyes. Study parameters consisted of graded symptoms and signs, and serial impression cytologies. A total of 161 patients were enrolled in the study. Of these patients, 116 who had a minimum of 4 to 8 months of follow-up qualified for final statistical analysis. These patients were classified into two major groups: (1) keratoconjunctivitis sicca (KCS) and (2) conjunctival cicatricial diseases (Stevens-Johnson syndrome, inactive pemphigoid, radiation-induced dry eye, drug-induced pseudopemphigoid, and toxic epidermal necrolysis). An analysis of adjusted mean changes for KCS patients showed no statistically significant differences between active drug and placebo. A similar analysis of patients with conjunctival cicatricial diseases indicated a statistically significant (P less than 0.05) reversal of conjunctival keratinization in the temporal bulbar site after treatment with active drug, however, clinical symptoms and signs showed no significant improvement with active drug relative to placebo. Side effects were limited to blepharoconjunctivitis and were reversible upon tapering or stopping the drug.