Purpose To investigate the in vitro effect of pH, osmolarity, solvent, and light interaction on currently used and novel dyes to minimize dye-related retinal toxicity. Design Laboratory investigation. Methods Retinal pigment epithelium (RPE) human cells (ARPE-19) were exposed for 10 minutes to different pH solutions (4, 5, 6, 7, 7.5, 8, and 9) and glucose solutions (2.5%, 5.0%, 10%, 20%, 40%, and 50%) with osmolarity from 142 to 2530 mOsm, with and without 0.5 mg/mL trypan blue. R28 cells were also incubated with glucose (150, 310, and 1000 mOsm) and mannitol used as an osmotic control agent in both experiments. Dye-light interaction was assessed by incubating ARPE-19 for 10 minutes with trypan blue, brilliant blue, bromophenol blue, fast green, light green, or indigo carmine (0.05 mg/mL diluted in balanced saline solution) in the presence of high-brightness xenon and mercury vapor light sources. Results Solutions with nonphysiologic pH, below 7 and above 7.5, proved to be remarkably toxic to RPE cells with or without trypan blue. Also, all glucose solutions were deleterious to RPE ( P < .001) even in iso-osmolar range. No harmful effect was found with mannitol solutions. Among the dyes tested, only light green and fast green were toxic to ARPE-19 ( P < .001). Light exposure did not increase RPE toxicity either with xenon light or mercury vapor lamp. Conclusions Solutions containing glucose as a dye solvent or nonphysiologic pH should be used with care in surgical situations where the RPE is exposed. Light exposure under present assay conditions did not increase the RPE toxicity.