OBJECTIVE To investigate the toxic effects of triamcinolone acetonide (TA) suspensions on human retinal pigment epithelial (RPE) cells. METHODS Cultured human RPE cells were exposed for up to 2 hours to one of seven solutions: control (balanced salt solution, BSS; Alcon Laboratories, Ft. Worth TX), commercial TA suspension (cTA), cTA from which the vehicle (which contains the preservative benzyl alcohol) had been removed (vehicle-removed TA, -vTA), vehicle of the cTA (V), or a 1:10 dilution (in BSS; Alcon) of cTA, -vTA or V. Solution effects were evaluated by phase-contrast microscopy of cells stained in situ with trypan blue and in vitro by trypan blue exclusion assay. RPE cell function was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. The mechanism of TA toxicity was studied by acridine orange-ethidium bromide staining and epifluorescence microscopy, and ultrastructural changes were examined by transmission electron microscopy (TEM). RESULTS The effects of vehicle-removed solutions (-vTA and 1:10 -vTA) were similar to those of the control solution. Exposure for 1 hour or longer to a vehicle-containing solution (cTA and V) resulted in similar and significant degrees of cell damage that were dose and time dependent. The major mechanism of cell death was necrosis, and the early ultrastructural change was swelling of organelles in the cytoplasm. CONCLUSIONS Preserved commercial TA suspensions damaged human RPE cells, but vehicle-free solutions did not. The authors suggest removing the vehicle as completely as possible from TA solutions before they are administered intravitreally. Furthermore, they recommend that a commercial formulation of preservative-free TA suspension be made available for intraocular use.