AIMS: The choice of a prophylactic antibiotic for cataract surgery is dependent on its antibacterial activity and tissue penetration. The influence of the route and timing of administration of cefuroxime on its intraocular concentrations was examined. METHODS: 120 patients were recruited before cataract surgery into a prospective trial to compare the anterior chamber concentration of cefuroxime at a fixed time after administration by three routes. In a further 110 patients, the interval before sampling was varied in order to permit an examination of the kinetics of penetration. In another 10 patients, cefuroxime was given topically at the completion of surgery to assess the effect of a corneal wound on aqueous penetration. Cefuroxime concentrations were measured by high performance liquid chromatography on 0.2 ml samples of aqueous aspirated from the anterior chamber. Mean aqueous concentrations of cefuroxime for each group were compared using Student's t test. RESULTS: After 25 mg cefuroxime, mean aqueous concentrations increased in the order forniceal (< 0.1 microgram/ml) < topical (0.18 microgram/ml) < subconjunctival (2.31 microgram/ml) when sampled 12-24 minutes after administration. Aqueous concentrations of cefuroxime reached a peak between 80 and 110 minutes after both forniceal and peribulbar injection but were still rising at this time after subconjunctival injection. Topical application of 12.5 mg cefuroxime to eyes with a 10 mm corneal wound resulted in a mean aqueous concentration of 9.34 micrograms/ml. CONCLUSION: In the intact eye, only sub-conjunctival injection resulted in clinically significant aqueous concentrations of cefuroxime (> 1 microgram/ml) between 12 and 24 minutes after administration. For all routes, maximal aqueous concentrations were delayed by at least 80 minutes from administration. In the presence of a corneal wound, high aqueous levels of cefuroxime were rapidly attained after topical application.