We investigated the dose-response and reproducibility of the intraocular pressure-lowering effect of MK-927 in ocular hypertensive patients. Patients were enrolled until at least 8 "marked responders" (peak reduction in intraocular pressure comparing the MK-927-treated eye with the placebo-treated eye greater than or equal to 6 mm Hg) and 7 "mild responders" (peak reduction in intraocular pressure comparing the MK-927-treated eye with the placebo-treated eye less than or equal to 3 mm Hg) were identified. In part A, 27 patients received one drop of 2% MK-927 in one eye (baseline mean +/- SEM intraocular pressure, 28.0 +/- 1.0 mm Hg) and placebo in the contralateral eye. Intraocular pressure was measured at baseline and 1, 2, 3, 4, and 6 hours. Maximum reduction in intraocular pressure was 4.0 +/- 0.8 mm Hg at 3 hours, with a duration of 4 hours. Ten patients were identified as marked responders and 7 as mild responders. In part B, 8 of the marked responders entered a four-period crossover study and received 2%, 0.5%, and 0.125% MK-927 and placebo in the same treated eye as in part A, and placebo in the contralateral eye. The 7 mild responders in part C received 2% MK-927 in a similar fashion as in part A. MK-927 in concentrations of 0.125% and 0.5% had little or no effect on intraocular pressure in patients with a marked response to 2% MK-927. Within-patient variability in peak response to single doses of 2% MK-927 was substantial (coefficient of variation of 0.3 and 0.5 for marked responder and mild responder groups, respectively.