Effect of Intraocular Pressure Control on Visual Field Progression in the HORIZON Trial

To analyze the effect of intraocular pressure (IOP) characteristics on the 5-year rate of visual field mean deviation (MD) progression in the HORIZON trial.

Post-hoc analysis of data from a randomized clinical trial.

Patients with glaucoma and visually significant cataract randomized to cataract surgery (CS) with or without the implant of the Hydrus Microstent (HMS).

One eye of each participant with at least 3 reliable visual fields (false positive errors ≤15%) over at least 1 year was included. We used a published Bayesian hierarchical model, designed to minimize the effect of perimetric noise and learning. We compared the mean rate of MD progression between the 2 arms and measured the effect of postoperative mean intraocular pressure (MIOP), peak washed-out intraocular pressure (WO-PIOP), mean washed-out intraocular pressure (WO-MIOP), and day-1 IOP. These IOP metrics were mean-centered and standardized (divided by their standard deviation) to allow a direct comparison of their effect.

Mean rate of MD progression.

Data from 517 eyes (352 in the CS-HMS arm) were analyzed. The MD rate was -0.55 [-0.72,-0.39] dB/year (mean [95% credible intervals]) for the CS-only arm and -0.30 [-0.38,-0.22] dB/year for the CS-HMS arm (44.6% [18.0, 64.2] reduction, P = 0.004). Increase in all IOP metrics was significantly associated with a faster MD rate. MIOP, WO-MIOP, and WO-PIOP had similar proportional effects. Day-1 IOP had a considerably smaller effect. The difference in rate between the 2 arms remained significant after adjusting for MIOP (P = 0.002) but was largely eliminated after adjusting for WO-MIOP and WO-PIOP (P = 0.396). The MD rate in the CS arm remained significantly faster than CS-HMS (P = 0.001) after removing eyes with a day-1 IOP spike (IOP > 40 mmHg or 10 mmHg above the washed-out baseline IOP).

The MD rate was significantly slower in the CS-HMS arm in the HORIZON trial, compared to CS alone. This difference was not explained by day-1 IOP spikes or MIOP, but was largely accounted for by WO-IOP metrics.

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.